This four-year project aims to strengthen programs that promote entry into and success in biomedical research careers, particularly by underrepresented minority students. The research questions are: 1) how do program activities, particularly research participation and mentoring, influence: a) students' skills in science inquiry and teamwork, b) their beliefs in their abilities, and c) their stage-appropriate career outcomes; and 2) are these influences similar or different for minority and non-minority students. Students from high school, community college, undergraduate, and graduate levels will participate in the study. Three separate studies will test theoretical hypotheses. An Alumni Retrospective Survey will involve 700 high school program alumni, 500 undergraduate program alumni, 100 graduate program alumni, and 500 undergraduate and 100 graduate science alumni who did not participate in formal support programs. These responses, as well as qualitative information from program faculty and staff and from field observations, will provide the foundation for the design and testing of simulations that serve both as performance-based assessments of science inquiry and scientific teamwork skills and as educational experiences. In the COSMOS Simulation and Survey, 280 high school students recruited for COSMOS because of their high abilities in science and mathematics will participate in the simulations at the beginning and end of their four-week residential program. In the Undergraduate Simulation and Survey, 100 undergraduates involved in science support programs will participate in the simulations at the beginning and end of the 2006-07 academic year. Across these studies, approximately 54% of the sample will be females, and 44% from underrepresented minority groups. With the student as the unit of analysis, our primary dependent measures are intention to continue and actual continuation toward research careers, skills in scientific inquiry and teamwork, and beliefs about those skills. Data will be analyzed primarily through correlational and multiple regression procedures. [unreadable] [unreadable]